Skip to content
llm-speed
Leaderboard/M4 Max/Qwen3-Coder-30B-A3B-Instruct

Can an M4 Max run Qwen3-Coder-30B-A3B-Instruct?

llm-speed has not measured Qwen3-Coder-30B-A3B-Instruct on M4 Max itself yet, so there is no llm-speed verdict. 4 community-submitted runs exist plus 1 with a non-default setting, listed separately (fastest 94.5 tok/s at Q4_K_M on ollama); community-submitted, not measured by llm-speed, so they give no verdict.

5 signed runs (4 default settings, 1 Ollama tag qwen3-coder-bench-32k) on llm-speed. Single-request decode speed in tokens per second; the short-chat workload unless noted. All 5 runs are community-submitted (signed on the submitter's own machine, not measured by llm-speed, and not used for llm-speed verdicts). Verdicts and headline numbers use only llm-speed's own runs and fit tests.

Measured tok/s by quant and runtime

QuantRuntimeBuildDecode tok/s (best run)WorkloadRunsSource
Q4_K_Mollama0.30.1194.5chat-short4Community-submitted
Q4_K_MOllama tag qwen3-coder-bench-32k (counted separately)ollama0.30.11113.3chat-short1Community-submitted

llm-speed has not measured this pair itself, so there is no llm-speed verdict or fastest quant. The rows above are community-submitted, not measured by llm-speed.

VRAM: not recorded for these runs. The VRAM calculator gives an estimate (predicted, not measured).

Every measured row

RunDate (UTC)WorkloadQuantRuntime / buildPrompt tokensOutput tokensDecode tok/sPrefill tok/sTTFT msVRAM GiBSource
r_roktphpc--82026-06-27chat-shortOllama tag qwen3-coder-bench-32kQ4_K_Mollama · 0.30.11110256113.384.11308–Community-submitted
r_r0di2hkku1h2026-06-27chat-shortQ4_K_Mollama · 0.30.1111025694.565.11691–Community-submitted
r_txvkjgoyyxs2026-06-27chat-shortQ4_K_Mollama · 0.30.1111025693.213.38270–Community-submitted
r_40h1dznuznk2026-06-27chat-shortQ4_K_Mollama · 0.30.1111025692.7719.0153–Community-submitted
r_1o8q4lhgj882026-06-27chat-shortQ4_K_Mollama · 0.30.1111025655.5675.2163–Community-submitted
r_roktphpc--82026-06-27chat-longOllama tag qwen3-coder-bench-32kQ4_K_Mollama · 0.30.11314757197.41342.82344–Community-submitted
r_r0di2hkku1h2026-06-27chat-longQ4_K_Mollama · 0.30.11314757497.41397.32252–Community-submitted
r_40h1dznuznk2026-06-27chat-longQ4_K_Mollama · 0.30.11314763494.61396.52253–Community-submitted
r_1o8q4lhgj882026-06-27chat-longQ4_K_Mollama · 0.30.11314767957.928422.4111–Community-submitted
r_roktphpc--82026-06-27agent-traceOllama tag qwen3-coder-bench-32kQ4_K_Mollama · 0.30.113247840103.23371.6477–Community-submitted
r_r0di2hkku1h2026-06-27concurrent-decodeQ4_K_Mollama · 0.30.111024256109.9–––Community-submitted
r_roktphpc--82026-06-27concurrent-decodeOllama tag qwen3-coder-bench-32kQ4_K_Mollama · 0.30.111024256109.7–––Community-submitted
r_40h1dznuznk2026-06-27concurrent-decodeQ4_K_Mollama · 0.30.111024256109.1–––Community-submitted
r_1o8q4lhgj882026-06-27concurrent-decodeQ4_K_Mollama · 0.30.11102425667.0–––Community-submitted

Questions

Can an M4 Max run Qwen3-Coder-30B-A3B-Instruct?

llm-speed has not measured Qwen3-Coder-30B-A3B-Instruct on M4 Max itself yet, so there is no llm-speed verdict. 4 community-submitted runs exist plus 1 with a non-default setting, listed separately (fastest 94.5 tok/s at Q4_K_M on ollama); community-submitted, not measured by llm-speed, so they give no verdict.

What did community runs measure for Qwen3-Coder-30B-A3B-Instruct on an M4 Max?

5 community-submitted runs (community-submitted, not measured by llm-speed): Q4_K_M on ollama: 94.5 tok/s; Q4_K_M (Ollama tag qwen3-coder-bench-32k) on ollama: 113.3 tok/s. They are listed for reference and give no llm-speed verdict.

Sources and related pages